Chemical behaviour of trace pollutants during roof and road runoff

Figure 1 : Plot of Henry's law constant H against octanol-water partition coefficient kow (fl '"M (PAH) <202, ... '" M (PAH)

>202 9 mol-1,

0 '"

.BaA ...

... BghiP

o

_\ -Ig kow .4-NP

i -2 i

-3 'V y-HCH

1

·4

1

-5

i -7 -6 +lgH +1

0

-,

·2

·3

·4

·5

·6 -7

!

-.

Reimer HERRMANN, Joachim DAUB, Jürgen FORSTER and Thomas STR/EBEL

Chair of Hydr%gy, University of Bayreuth, 95440 Bayreuth, Germany

CHEMICAL BEHAVIOUR OF TRACE POLLUTANTS DURING ROOF AND ROAD RUNOFF

The fates of ionic and nonionic organic compounds and trace metals during roof and road runoff are sensitive to their distri- bution between sorption onto roof and road material and sus- pended sediments on the one hand and dissolution on the other hand. Using field data of runoff, suspended sediment concen- tration and the chemical state of various trace pollutants, we try to explain the factors governing the chemodynamics and the transport behaviour during roof and road runoff.

1. AB5TRACT

2. INTRODUCTION

We will transfer the reasoning of CHIOU et al. (1985) about the sorption of nonionic organic compounds onto soil humus and minerais to the understanding of chemodynamic behaviour of nonionic organic pollutants on roofs and roads. We extend this reasoning further to hydrophobic ionizable organic chemicals (e.g. phenols) and mixtures of water and organic sol vents (e.g.

DOC) (SAWHNEY 1989, RAO et al. 1989). We should expect that wet surfaces of roof and road material and suspended sediments, both coated with organic material, would linearly, and without competition belween different solutes, uptake nonionic organic compounds from water by means of solute partitioning into organic matter. However, decreasing with in- creasing dipole interaction of water with minerais, nonionic or- ganics will be adsorbed on the minerais' surfaces by entropy driven processes. This process leads to competition between the various adsorbates in a mixture.

Strong vapour phase sorption by dry surfaces of roads, roofs and sediments can be explained by minerai adsorption which holds commanding influence of adsorption onto surfaces.

During wetting of surfaces, part of the previously adsorbed nonionic organic chemicals will desorb, others will partition into the organic material. However, the uptake of organic chemicals by organic coatings on wet roofs, roads and sediments is small in relation to organic solvents. In case of low contents of organic matter on road and roof surfaces, we should expect a change of activity of organic chemicals during wetting and drying cycles, because after wetting of the surface, water will displace the nonionic organic chemical from the minerai surfaces.

As an example of hydrophobic ionizable hydrocarbons SAWHNEY (1989) discusses the sorptive behaviour of phenols.

Partition coefficients predicted from water solubility or water- octanol partitioning are lower than those measured in sediment- water systems. He explains this discrepancy by sorption pro- cesses of phenols by soils and sediments which are related to specific interactions, such as by H-bonds (s. fig. 1), rather than by general hydrophobic mechanisms. Furthermore, phenolic compounds with a pka lower than the pH of roof or road runoff will dissociate, and the resulting phenolate has a much greater solubility than the protonated form.

Dissolved organic compounds act as cosolvents, thus de- creasing sorption of organic chemicals on surfaces and de- creasing the sorption coefficient. The result is a greater mobility of the organic che mi cal as a dissolved species within the runoff.

L1SS and SLATER (1974) describe the use of a Iwo layer model to estimate the flux of various gases across the air-water interface. The total resistance on a liquid phase basis (Ki') depends on the exchange constants k_land k of the individual phases and the dimensionless Henry's law

c~nstant

(H):

Ki' = ki' + (Hkgt '

Because of high values of H of ail the volatile organic hy- drocarbons we investigated, the Iiquid phase resistance con- trolls the exchange (ki' » H kg').

We plotted the distribution coefficient of the air-water system (Ig HENRY's law coefficient) against the octanol-water partition coefficient (Ig k_{ow )}in fig. 1 for a few trace organics : the volatility increases upward and the sorption increases to the left, thus the capability of escaping the aqueous phase decreases towards the lower right.

We will select examples for the dominant chemodynamic and transport processes which affect the behaviour of organic chem- icals during roof and road washoff. However, since at any time no single process acts alone, real world behaviour of organic chemicals cannot definitely be attributed to the single contrib- uting chemical processes.

The explanation of the chemodynamics of trace metals during roof and road runoff is severely limited as we do not know sufficiently of most of the influencing factors, e.g. the kinetics of dissolution and precipitation, the thermodynamic data and ligand concentrations, in order to estimate inorganic and organic metal complexes, and their saturation, and finally the data for computing sorption processes onto roof and road surfaces and onto suspended sediments. Especially the kinetics of partition- ing of trace metals into or out of suspended sediments or roof and road material is not weil known.

The sorption capacity of suspended sediments and roof and road surfaces primarily depends on the coating of the surfaces by amorphic oxides of Fe and Mn and by reactive organic carbon. Further, clay minerais and amorphic AI and Si oxides act as sorbents for trace metals or they form substrates which can be coated by Fe and Mn amorphic oxides or reactive organic carbon.

For a comprehensive treatise on the behaviour of trace metals in the .hydrocycle, we refer to the monograph of SALOMONS and FORSTNER (1984).

Analogous to soils, we can expect that surfaces of suspended

70 LA HOUILLE BLANCHE/W 112-1994

c(OOC)(mgll) clAt sortl.).C(fItdi".)(llg/~

6 O O - r - - - . - - - . - - - " T " 1 . 2 .5

0.8 2.0

1.0 1.5 1.0 c(At),c(Bghip) (llg/g)

"T""":.---,---....;...;..,;,..;...;...3.0

Figure 3: Temporal variation of specific concentrations of sorbed fluoranthene (Fit), benzo(ghi)perylene (BghiP), sus- pended sediments (55) and runoff (0) during road washoff (Bayreuth, Konigsallee, 14/07/1991).

Normalized Ireight sum

1.0 1.0

08 0.8

0.6 0.6

0.4 0.4

0.2 0.2

Figure 2: Road-runoff event : plot of normalized freight sums of fluoranthene (Fit), suspended sediments (55) and DOC against normalized flow-mass (Bayreuth, Konigsallee, 14/07/1991 ).

O~~...- - - , - - - " , -TI- - , . ,----r-- 0

o 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 normalized flow·mass

SS(mg/l) 1200 1000

Q(IIi') 800 ²⁰⁰ 600 ⁵⁽⁾ 400 ⁰⁰

200 0 0

400 500

3. STUDY AREA, SAMPLING AND CHEMICAL ANALYSIS

4. CHEMODYNAMICS AND TRANSPORT BEHAVIOUR OF TRACE ORGANIC COMPOUNDS

For this study we selected three road sections [HERRMANN et al. 1992, KERN et al. 1993] in the city of Bayreuth, five roofs within the university campus [FORSTER 1990] and five exper- imental roofs on top of a fiat roof also within the university campus [FORSTER 1991]. We developed automatic sampling systems [FORSTER 1991, STRIEBEL et al. 1993] which al- lowed to take samples and ta measure runoff, electric conduc- tivity and pH with high temporal resolution. We analyzed the organic micropollutants by means of solid phase and liquid/solid extraction, a cleanup step, GC and HPLC separation and ECD, MS, and fluorescence detection. Depending on their position within the speciation scheme, we measured trace metals by means of atomic absorption spectroscopy (AAS) and anodic stripping voltammetry (AS V) with or without pretreatment (e.g.

ion exchange, acid digestion under pressure or UV-radiation).

sediments and of roofs and roads present great numbers of surface coordination sites. Normally we observe a significant hysteresis between sorbing and desorbing trace metals result- ing from the metals partly entering the solid material or from diffusing into micro-pores or a larger endothermic energy need.

Thus, desorption needs more time to equilibrate than sorption does.

Besides the composition of the surfaces, the composition of the solute, too, influences the sorption behaviour of trace metals during roof and road runoff. The pH strongly influences the sorption of trace metals by hydrolysis or protonation of the ions and changes the charge of the surfaces.

Besides some exceptions, the complexation of trace metals by organic and inorganic ligands decreases the sorption onto surfaces and increases the mobility of the trace metals.

We are not able to predict the results of competitive behaviour of a trace metal and constituents of the solute, which potentially can adsorb at the same surface sites like the trace metal under consideration, since we generally do not know the composition of the solute and of the surfaces. Though the ionic strength affects trace metal sorption to a lesser degree than proton activity, complexation or competition, it influences metal ion activity in solution and the surface charges. Further it influences the formation of aggregates and thus sorption or desorption by increasing or decreasing the active surface area of the solid phase.

4.2. SORPTION PROCESSES

A plot of normalized freight sums against normalized flow mass (fig. 2) provides evidence that, during road washoff, fluoran- thene sorbed by suspended sediments behaves like the sus- pended sediments themselves, whereas the behaviour of dis- solved fluoranthene can be compared with that of DOC.

However, at times with intensive runoff, coarse suspended sediments are eroded, which have a lower ratio of organic coatings to minerai surfaces and thus a lower ove rail sorption capacity. As a result, the specific adsorption declines (s. fig. 3).

The combination of time series of temporal variations of dis- salved and sorbed fluoranthene with that of DOC indicates that the ratio of dissolved to sorbed fluoranthene during road washoff can at times be considerately higher than one would predict from the high k_ow(s. fig. 1). We explain this behaviour by cosolvation by nonpolar films of minerai oil, dissolved low molecular organic

Figure 4: Temporal variation of dissolved (diss.), sorbed (sorb.) fluoranthene (Fit) and DOC during road washoff (Bayreuth, Konigsallee, 14/07/1991).

chemicals, nitrophenols or other organic solvents (fig. 4). We observed this behaviour in particular with nonionic organic chemicals with very high sorption coefficient: during an event with c(DOC) = 2 - 10 mg/l, the ratio roof dissolved fluoran- thene/BghiP was r = 7, however at an another event with c(DOC) = 27 - 545 mg/I the same ratio decreased down to r= 1.

During dry weather, y-HCH (Lindane) deposits on roofs and roads in the gaseous phase. During wetting of the surface, this organic compound, which is only weakly adsorbed by minerais (Ig k_d < 2) and has a solubility of s= 1511mol/l, will readily

1/2-1994

O+--.,.---.,.---.,.---.,.---""'T"""'"

1.5 2,0

spec.mass -llow(1m²) ... rain

_ concrete tiles north o-aconcrete tiles south ... zinc sheets north .,..., zinc sheets south

Figure 7: Temporal variation of concentrations of benzo(a)an- thracene (BaA) during roof washoff on north and south sides.

c(Ba A),pmoll·1

partition coefficients k_{ow '} volatilè chlorinated hydrocarbons (s_

fig. 1) tend to easily equilibrate with the air, do not significantly deposit in vapour phase on road and roof surfaces, and do not significantly sorb onto minerai or organic surfaces. Thus, the concentrations of this class of nonionic organic hydrocarbons in roof and road washoff are governed by their concentrations in rainfall, and as a result the deviation of their normalized freight sum from the specific f1ow-mass is very low.

Comparing the concentrations of benzo(a)anthracene (s. fig. 7) in roof washoff from a tin roof and a concrete tile roof from north and south sides we observe that there is no significant differ- ence which could be interpreted to be caused by photochemical breakdown. The apparent differences which change during the event may be caused by different hydrodynamics and different dry deposition in luff and lee as weil as different residual mass left over from previous rainfall. ZARIFIOU et al. (1984) warn that great care must be taken to avoid experimental artifacts when dealing with photochemical reactions.

. 4.3. PHOTOCHEMICAL REACTIONS

1.5

o----arainfall }

. - - pantiles 20./21.6.1990

~fibre cement ... rainfall 116.8.1990

_ fibre cement

0.5 1.0

o

0.1 0.2 0.3 0.4

2.0 2.5₂

normalized runolt(1m- ) Figure 5: Concentrations of y-HCH during roof washoff from pantiles and fibre cement.

c(y -HCH), nmol1¹

0 . 5 , T - . . . . ; - . - - - .

desorb, and during the course of a runoff event will adopt concentrations of the feeding rainfalL Thus, at onset of rain, roofs with a higher inner minerai surface, like fibre cement, exhibit a considerable higher concentration of y-HCH in surface roof washoff than does the rain (s. fig. 5).

Because of their high solubility (s. fig. 1), nitrophenols .as an example of hydrophobic ionizable hydrocarbons are easily mo- bilized during road washoff. Nitrophenols with intermolecular H-bonds or pk_a < pH like 4-nitrophenol in our example (pk_a= 7,1) show a marked first flush effect (s. fig. 6). Contrary, the transport of nitrophenols adsorbed onto suspended sedi- ments is in most cases below the limit of determination and in general negligible. During times of intensive rainfall the con- centrations of nitrophenols decrease in road washoff caused by dilution and lower contribution by dissolution. Near the end of a rain event, however, concentrations in washoff are controlled by concentrations in rainfalL Further, during summer events with low runoff, warm road surfaces cause evaporation with the effect of an increase of nitrophenol concentrations. During strong road washoff the more mobile 4-nitrophenol reaches a state of supply limit considerately earlier than the less soluble and mobile 2- nitrophenoL

4.2. TRANSPORT ACROSS THE WATER-AIR INTERFACE

Caused by high HENRY's law constants and low octanol-water

5. CHEMODYNAMICS AND TRANSPORT BEHAVIOUR OF TRACE METALS

5.1. SPECIES DISTRIBUTION: CHEMICAL

SPECIATION .

Using two examples of chemical speciation, a meltwater and a rain runoff event (s. fig. 8), we want to emphasize that the species distribution, and thus sorption and transport behaviour, may change considerably between and within runoff events.

During the meltwater event with high ion concentration (X20

=

36.000_~S/cm),Cu was transported in equal shares in dissolved and adsorbed phases. The contribution of dissolved Cu was comparatively high. About a third of the dissolved Cu is ASV- labile bound, and this portion cannot be increased by UV- radiation. Equally, the portion of Chelex-Iabile metals does not increase. Thus, the not labile bound Cu species either exist as very stable inorganic complexes which do not disintegrate under the influence of a strong ion exchanger or as colloidal Cu species. The rain runoff resulted from a weak rain event, and thus the concentration of suspended sediments and as a result the transport of Cu in adsorbed phase was only low. Also, the

o ⁰

o

Figure 6: Temporal variations of dissolved 2-nitrophenol (2- NP), 4-nitrophenol (4-NP), electric conductivity (X 20) and pH during road washoff. For 4-NP, the concentrations in rain are also plotted (Bayreuth, Albrecht-Dürer-Stra~e.08/05/1990).

X20 (ilS/cm) c(Nitrophenole)( Il g/I)

pH

pH 8.0 30

300

1.5 25 1.0 20 200

15 4-NP(rain)

100 I!J 10

LA HOUILLE BLANCHE/N° 1/2-1994

5.2. METAL SPECIATION BV MEANS OF THERMODVNAMIC EQUILIBRIUM

Figure 8: Chemical speciation of Cu in meltwater and rainwater runoff sampi es (Meranierring, Bayreuth).

OH 92%

Law contamination Highcontamination

5045%

22 20 18 24

16 14 12

1

:1~

C(Cu).~g1·1 rain runotf 26

!~!

l~l~~~1rain runoff

c-tot ::-12 ....m

ASV-Iabile bound Cu was only low, but could be considerably increased by UV-radiation: a clear indication of ASV-stable, organic complexing agents. As the Chelex-Iabile and the UV- ASV-Iabile portions nearly sum up to the concentration of total dissolved Cu, dissolved Cu is practically completely exchange- able, but not existing in an extremely labile form.

50 15

meltwater runoH

125

100 150 200 C(Cu).~^{g l'}

115

Figure 9: Thermodynamic speciation of Cu, Pb and Cd in a road runoff with low and high contamination (Meranierring, Bay- reuth).

roofs with different mate rials and si op es between 15° and 30°

and one fiat roof. During the runoff event presented (s. fig. 10) the mean concentrations of suspended sediments in runoff related to that of rain c(ss, rain)=17 mg/dm³were : tarfeld roof

=258 %, pantile roof=211 %, fibre cement=281 %, zinc sheet roof=399 % and fiat gravel roof=26 %, resp. During the first 0,251/m² of runoff, the concentrations of Pb adsorbed onto suspended sediments differ greatly. We explain this by different amounts of dry deposition before onset of rain and different washoff efficiency of suspended sediments caused by roof roughness and resistance to flow rather than pH-variation which wasD. pH < 0,5 during the event.

Computation (s. fig. 9) reveals that the separation into free dissolved and inorganic metai complexes cannot be related to the general state of contamination of a runoff sample. Rather, the separation depends on the availability of the complexing ligands. Cu tends to complex with chloride and sulfate as weil as with hydroxide and carbonate. As a result, we observe at the

" high contamination case" a significant complexation by chlo- ride and sulfate caused by the availability of these anions and a low pH = 6,2. This low pH caused a low carbonate and hydroxide complexation and a high concentration of free Cu. On the other hand, low chloride and sulfate concentrations and a high pH=7,6 at the«low contamination case" shift the species distribution to a dominant carbonate and hydroxy-complexation.

The species distribution of Pb is caused by the same gov- erning factors. However, the distribution is reduced to sulfate and hydroxide complexes, which are related to contamination and pH, resp., and to free dissolved Pb.

ln case of Cd, only chloride and sulfate complexes and free Cd are of importance. Thus, the species distribution is clearly related to the availability of chloride and sulfate caused by contamination.

ln most cases concentrations of the three metals are below saturation. Only in the« low contamination case» we computed a weak oversaturation (SI=+0,1) of free Cu against CuO. This oversaturation may be a computational or experimental artefact caused by neglect of monovalent Cu ions, neglect of organic ligands or inclusion of colloidal Cu which passed the filter. Thus, we cannot refute the hypothesis, that the concentrations of dissolved trace metals are governed by competitive reactions between adsorption equilibria related to adsorption sites and dissolved complexes.

5.3. INFLUENCE OF DIFFERENT ROOF

MATERIALS ON TRANSPORT BEHAVIOUR

1250 1000 150 500 250

025 015 100

... tar felt roof ... pantile roof ... Etemlt roof o--a zinc sheet roof ... fiat graver roof

125 150 1.15 2.00

standardized runoM (1 m")

FORSTER (1990) investigated the pollutant load of runoff of five

Figure 10: Concentrations of adsorbed Pb in roof runoff (18/05/1988).

73

Figure 12: Temporal variation of concentrations of Pb (ad- sorbed onto particles>12 Ilm, < 12 J-lm and dissolved) (Bay- reuth, Konigsallee 14/07/1991).

Pb.diss.

Cu C(Pb,diss.), l'gl"

14 12 10

200

100 C(Pb.ads.),I'Ql"

400

300

C(Cu,Pb,ads.~12I'm).I'Q gol 1400

1200 1000 800 600 400 21JO

5.4. CHEMODYNAMICS OF Cu AND Pb DU RING A RUNOFF EVENT

Figs. 11 to 13 depict the temporal variations of various Cu and Pb fractions. We observe a weil marked coherence between the temporal variation of coarse suspended sediments and the concentration of Cu adsorbed onto coarse suspended sedi- ments, but with a temporal reversai of the height of the two peaks. It is evident that during peak washoff of coarse sus- pended sediments, these were only weakly contaminated. Even more clearly, this can be shown by means of the temporal variation of the specific adsorption of Pb and Cu onto coarse suspended sediments (s. fig. 13) where alter 3 h at the time of peak washoff we observe a minimum of concentration. In gen- eral, the specific concentrations of Cu decreased : particles with higher contamination are preferably washed off. Further, at the time of the second peak of washoff of coarse sediments, we do not observe a change in specific concentration. Thus, at this time a higher washoff of particl es leads to a proportional in- crease in particle bound copper. The washoff of Cu adsorbed onto fine suspended sediments follows the washoff of fine sus- pended sediments, with, however a distinctly higher specific concentration at the beginning. The concentration of dissolved Cu decreases with time in the same way as CI - and electric conductivity which are indicators of conservative dissolved sub- stances.

8 (h) O , ; - - - , - - - - ; - - r - _ - r - _ , - _ - - . _..._-i'li=me

o

REFERENCES

Figure 13: Temporal variation of concentration of Pb and Cu on suspended sediments > 12 Ilm (Bayreuth, Konigsallee 14/07/1991 ),

ACKNOWLEDGEMENT

The specific concentrations of trace metals (mass trace metal- /mass of sediment) adsorbed onto fine particles are greater than for coarse particles. However, in respect to the far greater surface area this difference in concentration is surprisingly low.

We thank the German Research Council and the Federal Min- istry of Research and Technology for support of this work. The discussion about the chemodynamics and transport behaviour of trace organics will be presented at the 6th I.C.U.S.D. in s'eptember 1993, Niagara Falls. .

500 400 600

300 200

-2 v-.".,,~--.::::::,+c::::::::._1100 0.45 <SS<12~m

20 40 30 50

10

ss>12~m

O+--,-":""--,---,---r--,...--r--r--.,...Lo

o 1 2 8 (h)lime

C(Cu.ads..>12),~g1,1

C(Cu.diss.).~gl" ClCu.ads..c12 ).l'gl"

300 10

21JO

100

C(CI' l.mg fI .1 .,

C(0.45 cSS c12~m).mg1 C(SS>12~m)mg1

6oIT---:..--:..~700

O+--'---'---'---r--r--r--r--~O

o ~~

Figure 11 : Temporal variation of concentrations CI-, sus- pended sediments (SS > 12 J-lm, 0.45 < SS < 12 J-lm), top., Cu adsorbed onto particles (> 12 Ilm and < 12 Ilm) and dissolved Cu (Bayreuth, Konigsallee 14/07/1991).

The temporal variation of Pb adsorbed onto coarse sus- pended sediments resembles that of Cu. However, we do not observe a general decrease of specific concentrations such as with Cu. Dissolved Pb and Pb adsorbed onto fine particles show a temporal variation similar to that of CI - or electric conduc- tivity and fine suspended sediments, resp. Thus, these temporal variations are similar to the resp. ones of Cu.

CHIOU C.T., SHOUP T.D., PORTER P.E., 1985 - Mechanistic roles of soil humus and minerais in the sorption of nonionic organic compounds from aqueous and organic solutions. Org.

Geochem., n° 8, pp. 9-14.

FORSTER J., 1990 - Roof runoff : a source of pollutants in urban storm drainage systems. Proc. 5th Int. Conf. Urban Storm Drainage, Osaka, pp. 469-474.

FORStER J., 1991 - Roof runoff pollution. ln: Grottker M., Schilling W. (Eds.) - Hydrological and pollutional aspects of stormwater infiltration. Proc. Sec. Europ. Junior Scient. Work- shop, Kastanienbaum, pp. 145-158.

HERRMANN R., DAUB J., STRIEBEL Th., 1992 - Qualitative Beurteilung der Niederschlagsabflüsse. In : Hahn H. H., Xantho- poulos C. - Schadstoffe im Regenabtluj3JI, Institut für Siedlungswasserwirtschaft, Karlsruhe, n° 64, pp. 115-155.

KERN U., WUST W., DAUB J., STRIEBEL Th., HERRMANN R.,

1992 - Abspülverhalten von Schwermetallen und organischen Mikroschadstoffen imStra~enabflu~. gwf - Wasser/Abwasser, n° 133, pp. 567-574.

L1SS P.S., SLATER P.G., 1974 - Flux of gases across the air-sea interface. Nature, n° 247, pp. 181-184.

RAO P.S.C., LEE L.S., NKEDI-KlllA P., YALKOWSKY S.H., 1989 - Sorption and transport of organic pollutants at waste disposai sites.

ln: GERSTL l., CHEN Y., MINGELGRIN U., YARON B. - Toxic organic chemicals in porous media - Springer, Heidel- berg, pp. 176-192.

SALOMONS W., FORSTNER U., 1984 - Metals in the hydro- cycle. Springer, Heidelberg.

SAWHNEY B.L., 1989 - Movement of organic chemicals through landfill and hazardous waste disposai sites. In : Sawh- ney B.L., Brown K. (Eds.) - Reactions and movements of organic chemicals in soils, Madison: Soil Sc. Soc. Am., pp. 447- 474.

STRIEBEL Th., DAUB J., HERRMANN R., 1993 - A sampling device for measuring physical and chemical characteristics of urban street runoff. The 4th International Symposium^«Highway Pollution,., Madrid (in print).

lAFIRIOU O.C., JOUSSOT-DUBIEN J., lE PP R.G., llKA R.G., 1984 - Photochemistry of natural waters. Environ. Sci. Tech- nol., n° 18, pp. 358A-371 A.

DISCUSSION

M BACHOC (DIREN-Centre, Service de Bassin Loire-Bretagne) M BRIAT, pouvez-vous nous dire quelles sont les caractéristi- ques des événements 4, 5 et 6 que vous avez présentés et qui sont intéressants, non-seulement du point de vue de la grande proportion de pollution soluble pour un certain nombre de polluants métalliques, mais aussi parce que leur concentration en MES est de l'ordre de 25 mg par litre, ce qui est assez surprenant et rare?

M BRIAT

Les événements 5 et 6 avaient respectivement 22 et 26 mg/I de concentration moyenne en MES, ce qui est tout à fait excep- tionnel. Cela nous a surpris.

Les caractéristiques des précipitations sont les suivantes:

14 mm pour l'événement 5 et 11 mm pour l'événement 6. Ce sont des pluies que nous pouvons considérer comme moyen- nes.

M DESBORDES

En ce qui concerne les échantillons, je constate que l'on conti- nue à échantillonner à pas de temps constant. C'est peut-être plus simple pour faire fonctionner les appareils, mais sur le plan de la représentativité des mesures, ce n'est pas nécessairement justifié. Alors, pourquoi ne développe-t-on pas des techniques de prélèvement qui sont asservies aux mesures des volumes passés et faisant des prélèvements en fonction du volume par exemple?

M BRIAT

Il nous a déjà fallu élaborer la courbe hauteur-débit et on ne connaissait les débits que par une courbe hauteur-débit qui n'était pas étalonnée au moment de l'installation. Ensuite, on a mesuré ces débits par cette courbe hauteur-débit et on a pu quand même recalculer les concentrations moyennes. On pensait qu'un temps de 10 minutes était quand même satisfai- sant. On avait peut-être peur d'avoir des échantillons soit trop

rapprochés, soit trop éloignés dans le temps les uns des autres si les débits variaient trop, mais on a pensé qu'avec un pré- lèvement d'échantillons toutes les 10 minutes on pouvait avoir, vu les vitesses de variation des paramètres, un échantillonage satisfaisant, du moins pendant la première heure et demie.

M DESBORDES

En ce qui concerne la mesure des hydrocarbures, M DUCREUX, vous avez donné des informations sur des recher- ches effectuées sur les précipitations. Le pluviomètre que vous avez utilisé était-il un pluviomètre fermé par temps sec ou ouvert ?

M DUCREUX

Nous avons utilisé deux types de pluviomètres: des pluviomè- tres qui étaient en permanence ouverts et d'autres qui n'étaient découverts qu'au moment des temps de pluie. On peut obser- ver, mais cela n'apparait pas très nettement, que les concen- trations en hydrocarbures obtenues sont un peu plus fortes sur le pluviomètre en permanence découvert. En ce qui concerne la nature des hydrocarbures, on a pu mettre en évidence des résultats quasiment identiques, excepté toutefois l'apparition de normale paraffine à répartition impaire en fin de chromatogram- me, qui est caractéristique des cires cuticulaires que l'on rencontre chez les végétaux.

M GIERSCH

Dans le cas des rinçages de toitures, est-ce que les différentes études ont pu permettre d'obtenir des conclusions ou des orien- tations de traitement, lorsque l'on veut par exemple respecti- vement infiltrer ou stocker sur la parcelle, pour réutiliser l'eau de toiture pour des questions d'arrosage, etc. ?

M HERRMANN

This, in fact is a great problem and a great discussion in Germany at the moment. We, ourselves, proposed that you'lI use soakaways where you can infiltrate roof and road runoff, but one should give the first one or two minutes into the sewage plant, and only have the lower concentration at the end going into soakaways.

112-1994 75